Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 25 Oct 2020 21:39:26 -0700
From:      Vincent Milum Jr <freebsd-arm@darkain.com>
To:        "freebsd-arm@freebsd.org" <freebsd-arm@freebsd.org>
Subject:   Re: FreeBSD under VMware ESXi ARM Fling
Message-ID:  <CAOWUMWGnm7bCtEuaRB5xH7BnEDCXj2W6kK1KK2QnaxHS3Yti6A@mail.gmail.com>
In-Reply-To: <CAOWUMWEGNN%2BHrRjN8k%2B5Et=RyK-i3QVBCgRb_3=uvx6jgZWf%2Bw@mail.gmail.com>
References:  <CAOWUMWF-%2BRg2syHnPZN28i1TB0-jr3P_eNsnSTL_48pZGqWjSw@mail.gmail.com> <gkECxHG3AZsNQOlRwjvFTnG3WZNYizcF5LicaGDvV1_qzBWqtJW4kSNRSZ-GEI3J8GKEjifr7kq-xBnGb-SjjjQt33MgoLsSTTQUubYVyh8=@protonmail.com> <CAOWUMWGWECs2VVo9afE5KB5csAQGCEq-=d4EbSJP5kFxv2mg5w@mail.gmail.com> <MZwl7SSjlHcZjH-qrI8Y2N96hag_TxajCc0K-Lmk6dPB7G64Nn7xm2e9ciPi8t9lK7yzgonrcy2zchruftUdat-FzCTbiUKJ8xu0Zr8fED0=@protonmail.com> <CAOWUMWEXiiw-avXQd9GKKn=xLdgxVKuU=opZ67UfkwnQvJb0GQ@mail.gmail.com> <KhcHxcOB6jmtZ7YCIkdHDx9-W2Wgu5TeC5hHjTfJsHVIDw20HwR9bVUZ7oyFHAj7SwCQzbI1LMaKKt0YplQy1WCIo4fmaTGpv_MDlYS1Zus=@protonmail.com> <CAOWUMWEGNN%2BHrRjN8k%2B5Et=RyK-i3QVBCgRb_3=uvx6jgZWf%2Bw@mail.gmail.com>

next in thread | previous in thread | raw e-mail | index | archive | help
As a follow up, the VMware devs have been looking into the SMP issue for
FreeBSD support. First, they've identified one bug for sure on their end
that will be fixed in ESXi ARM Fling v1.2. After that fix, the boot process
was blocking later on.

>From Cypou:
"Got something! I'm not sure where is the bug (esxi or freebsd), but the
virtual gic does not record cpu0 enabling its SGIs (IPIs) interrupts. By
forcing that in the hypervisor, I got all 4 cores running happily"

Source:
https://twitter.com/cypou/status/1320578258740662273

So now it just needs to be determined if this bug is on the ESXi side or
the FreeBSD side, and then we should be good!

On Tue, Oct 20, 2020 at 11:47 AM Vincent Milum Jr <freebsd-arm@darkain.com>
wrote:

> Most of those instructions are just for "update the RPi's firmware" - if
> you're on a new enough firmware already, you can skip that entire section
> involving Raspberry Pi OS.
>
> After that, they just provide an installer image which can be used to
> install to either local storage, or remote network storage. It is being
> treated more like a typical x86 workflow rather than what we're more
> familiar with for ARM/SBC workflows.
>
> On Tue, Oct 20, 2020 at 11:40 AM Robert Crowston <crowston@protonmail.com>
> wrote:
>
>> > According to the ESXi devs, there is no SIO controller currently, and I
>> believe this is required for the virtual serial ports to work
>> > https://flings.vmware.com/esxi-arm-edition/bugs/1104
>>
>> I do think this a requirement. Indeed, a UART is detected in your dmesg:
>>
>> uart0: <16550 or compatible> mem 0x3fff03f8-0x3fff03ff irq 5 on ofwbus0
>> uart0: console (230400,n,8,1)
>>
>> Is it not exposed on the host?
>>
>> If we cannot debug, a verbose boot would at least add some colour. Could
>> you add boot_verbose="YES" to /etc/loader.conf?
>>
>> I took a look at the instructions for getting this up and running on my
>> own Pi. It is certainly a lot of instructions. It is a shame they do not
>> just provide a .img I could flash to an sdcard.
>>
>>     -- RHC.
>>
>>
>>



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAOWUMWGnm7bCtEuaRB5xH7BnEDCXj2W6kK1KK2QnaxHS3Yti6A>